Imaging Workup of Acuteand Occult Lower

Gastrointestinal Bleeding Trevor C. Morrison, MDa, Michael Wells, MDb,Jeff L. Fidler, MDb, Jorge A. Soto, MD, PhDa,*

KEYWORDS

� Gastrointestinal � Bleed � Acute � Occult � Lower � Radiology � Angiography � Enterography

KEY POINTS

� Lower gastrointestinal bleeding is defined as occurring distal to the ligament of Treitz and presentsas hematochezia, melena, or with anemia and positive fecal occult blood test.

� Imaging tests in the workup of acute lower gastrointestinal bleeding include computed tomography(CT) angiography, nuclear medicine scintigraphy, and conventional catheter angiography.

� Imaging tests in the workup of occult lower gastrointestinal bleeding include CT enterography andnuclear medicine Meckel scan.

INTRODUCTION

Lower gastrointestinal (GI) bleeding is a frequentcause for hospital admissions with an annual inci-dence of approximately 20 to 27 cases per100,000 persons in the United States.1 Morbidityand mortality vary according to the underlyingcause of the GI bleed, with reported mortalityrates of 2% to 20% for lower GI bleeding and ashigh as 40% for hemodynamically unstablepatients.2

Lower GI bleeding is defined as bleeding thatoccurs distal to the ligament of Treitz, with upperGI bleeding occurring proximally. Clinical presen-tations vary based on the source of the bleed andcause; however, acute lower GI bleeds typicallypresent with hematochezia, noting that second-ary to the cathartic effects of blood, a brisk upperGI bleed may present in a similar manner.3

Causes of lower GI bleeding may be anatomic,such as diverticulosis (33.5%); vascular, suchas hemorrhoids (22.5%), angioectasia, or

Disclosure Statement: Nothing to disclose.a Boston University Medical Center, 830 Harrison Ab Department of Radiology, Mayo Clinic, 200 First Street* Corresponding author.E-mail address: Jorge.Soto@bmc.org

Radiol Clin N Am 56 (2018) 791–804https://doi.org/10.1016/j.rcl.2018.04.0090033-8389/18/� 2018 Elsevier Inc. All rights reserved.

ischemia; neoplastic (12.7%); inflammatory aswith inflammatory bowel disease; or infectious.4

If the workup of the large bowel is negative,then patients are suspected of having a smallbowel bleed.

There are several classification schemes usedto describe lower GI bleeding related to theduration and severity of the bleed as well asthe results of upper and lower endoscopy/imag-ing. When correlating with the amount ofbleeding, lower GI bleeds can be categorizedas massive, moderate, or occult. Massivebleeding is defined by the passage of profusehematochezia with hemodynamic instability.Moderate bleeding reflects hematochezia in he-modynamically stable patients. Occult bleedingrefers to the presence of a positive fecal-occultblood test or iron deficiency anemia withoutanother identifiable source and without frankhematochezia.5 Obscure bleeding refersto patients who have recurrent bleeding afternegative endoscopic evaluation and advanced

venue, FGH 3rd Floor, Boston, MA 02118, USA;Southwest, Rochester, MN 55905, USA

radiologic.th

eclinics.com

Morrison et al792

radiologic assessment of the small bowel andcan be either acute or occult.2

WORKUP RECOMMENDATIONS FOR ACUTELOWER GASTROINTESTINAL BLEEDING

The workup of patients presenting with acutelower GI bleeding involves resuscitation, localizingthe site of bleeding, and intervention to stop thesource of the bleeding, as appropriate. The maintools of the workup include direct visualizationwith proctoscopy/colonoscopy and imaging withcomputed tomography angiography (CTA), nu-clear scintigraphy, or angiography. Although sur-gery was once a necessity for patients withongoing lower GI bleed, advanced techniques inendoscopy and angiography have improveddetection and treatment, with surgery nowreserved for cases in which more conservativemanagement has failed.6

The clinical presentation of patients during thetriaging process as well as the services availableat a hospital dictate the order and priority in whichtests are used in the workup of an active lower GIbleed.1 Any patient with hemodynamic instabilitymust first be resuscitated. If endoscopy is avail-able, it is generally the first test preformed; howev-er, there are significant limitations. The colon mustfirst be prepped in order to clear enteric contentsand blood, which could obscure the source ofbleeding. Even rapid bowel preparations take atleast several hours, which may not be possible inpatients with ongoing bleeding. Additionally, insome series, colonoscopy detects the source ofbleeding in only 13% to 40%.2 In cases wherebyemergent endoscopy is not indicated, patientswill typically be sent for an imaging study, suchas CTA, nuclear scintigraphy, or catheter angiog-raphy, depending on the local availability and clin-ical expertise. In patients who are clinically stableat presentation, more conservative managementis generally indicated, with many patients beingworked up with elective endoscopy.7

COMPUTED TOMOGRAPHY ANGIOGRAPHY

CTA has excellent sensitivity and specificity for theidentification and localization of acute GI bleeds. Ameta-analysis by Garcia-Blazquez and col-leagues8 reported a sensitivity of 85.2% and spec-ificity of 92.1% for the detection of acute, active GIbleeds. Yoon and colleagues9 used arterial phaseCTA in 26 patients with massive GI bleeding andreported an overall accuracy of 89%. CTA candetect bleeding rates as low as 0.3 mL/s inin vitro studies.10 Advantages of CTA include thatit is widely available, noninvasive, provides

excellent localization, and additionally can providethe cause of the GI bleed. CTA can diagnose diver-ticular disease, vascular abnormalities (such asangioectasia), tumors, colitis/enteritis, bowelischemia, and postoperative/iatrogenic causes.Additionally, the arterial phase imaging demon-strates vascular anatomy and any vascular vari-ants. This information can be observed on thesource images and reformatted using maximumintensity projections (MIPs) to provide valuableinformation for the subsequent mesentericangiography.2

Disadvantages of CTA include radiation dose,as it is usually done in 3 phases; however, im-provements in CT dose reduction have madethis less of an issue than in the past.11 CTAalso requires an intravenous (IV) contrast dose,which is relatively contraindicated in patientswith acute renal failure. As with all imaging tests,patients must be actively bleeding at the time ofthe scan.CTA for acute GI bleeding is typically preformed

as a triphasic study. Enteric contrast is notgiven as is critical to acquire the examination asquickly as possible in patients with active hemor-rhage. A precontrast examination is first acquiredusing low radiation dose settings. The purpose ofthe precontrast portion of the examination is toidentify any hyperdense enteric contents, suchas pills, residual oral contrast, hyperdense stool,and so forth, so they are not confused for contrastextravasation on later phases. Next, at least100 mL noniodinated contrast is administeredintravenously by a power injector at a rate of 4 to5 mL/s. Automated bolus technique is preferred,with arterial phase obtained 8 to 10 seconds afterthe attenuation coefficient in the proximal abdom-inal aorta reaches a threshold of 150 Hounsfieldunits (HU). Portal venous phase is then acquiredapproximately 50 seconds after start of the arterialphase.12

Computed Tomography Angiography:Findings

Active GI bleeding is manifest by contrast extrava-sation into the bowel lumen on CTA. The contrastextravasation appears as a blush on arterial phaseimaging, which propagates further down thebowel on the portal venous phase due to peri-stalsis11 (Figs. 1 and 2). If bleeding is arterial, ajet of contrast may be seen. Lower intensity bleedsare often better seen on the portal venous phase,as more time has been allowed for the contrastto accumulate.13 It is important to look at the pre-contrast and contrast-enhanced images side byside, as inherently hyperdense intraluminal

Fig. 1. CTA of active GI bleed. A 76-year-old woman with history of diverticulosis and recurrent lower GI bleedspresents with hematochezia. Axial precontrast (A) and arterial phase (B) images from a CTA examination demon-strate contrast extravasation into bowel lumen in the descending colon (white arrow), consistent with active GIbleeding. Precontrast images are critical to ensure that contrast extravasation is not confused for hyperdenseenteric contents. Subsequent colonoscopy revealed a diverticular bleed.

Acute and Occult Lower Gastrointestinal Bleeding 793

contents can mimic contrast extravasation. Insome cases, active bleeding with contrast extrav-asation is not identified, but there is visualization ofa clot at the site of bleeding. The clot should behyperdense compared with the enteric contentson precontrast images. HU should measureapproximately 30 to 45 HU for unclotted bloodand 40 to 70 HU for clotted blood.13 The area ofhighest HU should be at the origin of the bleed,which is called the sentinel clot.14

NUCLEAR SCINTIGRAPHY

Nuclear scintigraphy is most commonly performedcombining the patients’ own red blood cells

Fig. 2. CTA of active GI bleed. An 87-year-old woman withwith hematochezia. Axial arterial phase (A) and portal vention into the sigmoid colon adjacent to the anastomosis. Nimages. A subsequent sigmoidoscopy revealed ischemic co

tagged with a radiotracer technetium 99-m(Tc-99m). Autologous red blood cells are taggedin vitro with 20 to 30 mCi Tc-99m and then admin-istered as an IV bolus. Patients are placed supineunder the gamma camera and scanned fromthe xiphoid process through the pubic symphysis.Initial dynamic flow images are obtained at3 seconds each, for 1 minute, to show flow in theabdominal viscera. This imaging is followed bystatic images every minute continued for 90 mi-nutes. If patients have an episode of rebleeding,additional static images can then be obtained upto 24 hours after the initial injection. Although pa-tients can be reimaged up to 24 hours after injec-tion without rebleeding, it is often not useful, as

history of hemicolectomy due to colon cancer presentsous phase (B) images demonstrate contrast extravasa-ote that there is increasing extravasation on delayedlitis at the surgical anastomosis.

Fig. 3. Nuclear scintigraphy of active GI bleed. A76-year-old woman with history of diverticulosis andrecurrent lower GI bleeds presents with hematochezia(same patient as in Fig. 1A, B). A 2D planar imagedemonstrates increased radiotracer uptake in the re-gion of the splenic flexure of the colon (black arrow),consistent with active GI bleeding.

Morrison et al794

peristalsis can provide false localization of the siteof the bleed. When positive findings are seen, theexamination can be discontinued once the clini-cian is confident of the origin of the bleed.15

Nuclear scintigraphy is the most sensitive test toimage GI bleeding, as it can detect bleeding ratesas low as 0.05 mL/min.16 Other advantagesinclude that the test is noninvasive and that no pa-tient preparation is needed other than obtaining ablood sample.The main disadvantage of nuclear scintigraphy

is that it generally relies on 2-dimensiona (2D)planar imaging, and localization of the bleed canbe a significant limited factor. For example, aredundant transverse colonic loop in the pelviscould be confused for a small bowel loop ormore distal colonic bleed. Also, because of inter-mittent bleeding and noncontinuous imaging, theradiolabeled blood may have already moveddownstream from the bleeding site at the timethe image is acquired, thus, giving false localiza-tion of the origin of the bleed. Single-photon emis-sion CT/CT fusion imaging has shown promise toimprove localization by pairing a CT scan with3-dimensional radiotracer imaging; however, thisis not widely available.17

In many institutions, nuclear scintigraphy is per-formed before angiography, because of thehigher sensitivity for detection of active bleeding,with a negative examination potentially elimi-nating the need for the more invasive angio-graphic procedure.18 When the study is positive,studies have shown improved bleed detection insubsequent angiography.19 However, althoughnuclear scintigraphy is the most sensitive imagingstudy to detect GI bleeds, CT is far more widelyavailable, with faster acquisition, and improvedlocalization, making it the test of choice at manyinstitutions.

Nuclear Scintigraphy: Findings

A positive nuclear bleeding study will show focalincreased activity and radiotracer uptake at thesite of the bleeding (Fig. 3). Over time, with bowelperistalsis, the radiolabeled blood will continue tomove along the GI tract, which aids in localiza-tion. Observing the course the blood takes overtime enables the clinician to determine if theblood is following a large or small bowelpathway. Imaging should be continued until theclinician is confident that they can identify thesource of the bleed.

MESENTERIC ANGIOGRAPHY

Mesenteric angiography is typically reserved fortreatment of bleeding that was first diagnosed

on a previous CTA or nuclear scintigraphy. Se-lective mesenteric angiography can detectbleeding at a rate of 0.5 mL/min, an approxi-mately 10-fold higher rate than that detectedby nuclear scintigraphy.20 The main advantageof angiography is that it can be used to delivertreatment if a focus of active bleeding is identi-fied, such as embolization or vasopressin drip.Angiography also does not require bowel prepa-ration and provides excellent localization of thesite of bleeding.There are significant disadvantages of using

angiography as a first-line diagnostic method inGI bleeding. Angiography is invasive, may requirea large load of iodinated contrast, and the radiationdose can be high in complicated cases. Angiog-raphy is not as widely available as CT, and manyhospitals lack expertise. An associated complica-tion of angiography is embolization causing bowelischemia.For diagnosis, angiography is reserved for pa-

tients with life-threatening, ongoing lower GIbleeding, when there may be no time for a CTAor nuclear study. In these cases, emergent angiog-raphy is often performed as a last remaining optionbefore laparotomy. Typically, the superior mesen-teric artery is cannulated first, as most significantlower GI bleeds occur from the right colon. If nobleed is identified, the inferior mesenteric arteryis cannulated followed by the celiac axis. Digitalsubtraction angiography (DSA) is a techniquethat uses a computer algorithm to subtract a

Fig. 4. CTA and DSA of active GI bleed. A 65-year-old woman with recurrent lower GI bleeds, without identifiablesource despite extensive workup, presents with hematochezia. Axial image from a CTA (A) demonstrates active GIbleeding within the small bowel (white arrow). DSA of the superior mesenteric artery (SMA) (B) demonstratessmall focus of extravasation in the pelvis (black arrow). Note the embolization clips from prior treatment. DSAof the SMA (C) after embolization demonstrates no evidence of contrast extravasation (black arrow).

Acute and Occult Lower Gastrointestinal Bleeding 795

precontrast image from a postcontrast image, sothat only the blood vessels are visualized. Thistechnique should be used if available, as it isgenerally thought to be more sensitive than angi-ography alone in the detection of GI bleeds.21

Mesenteric Angiography: Findings

On DSA, a lower GI bleed is identified as an area ofcontrast extravasation outside of the normal con-fines of a blood vessel, which typically increasesover time (Fig. 4). After successful embolization,an additional angiogram can be done to showthat the bleeding has stopped. Angiography alsohas characteristic findings in angiodysplasia,which often appears as a vascular blush with earlyand persistent visualization of a draining vein21

(Fig. 5).

WORKUP RECOMMENDATIONS FOR OCCULTGASTROINTESTINAL BLEEDING

Occult GI bleeding is diagnosed when patientshave a positive fecal occult test or iron deficiencyanemia without another source and the absenceof visible enteric blood. Upper endoscopy and co-lonoscopy identify an upper GI source of occult GIbleeding in 29% to 56% and a lower GI source in20% to 30% of patients; synchronous lesions in

Fig. 5. CTA and DSA of active GI bleed. A 58-year-old womness. Axial CT in portal venous phase (A) demonstrates conlective DSA of the middle colic artery in the early (B) and mof contrast (black arrow). Patient was treated with coil emcolon diagnosed on previous colonoscopy.

the upper and lower tract are discovered in 1%to 17%.22

Patients with negative esophagogastroduode-noscopy and colonoscopy by definition have sus-pected small bowel bleeding (SBB). Small bowellesions are responsible for 5% to 10% of all pa-tients presenting with GI bleeding.23 Video capsuleendoscopy (VCE), CT enterography (CTE), anddouble balloon endoscopy (DBE) can all be usedto assess the small bowel. VCE is recommendedby gastroenterology societies as the first-line testbecause of its high yield (40%–60%) of positivefindings after negative endoscopy; however, thetest is limited by low specificity, the potential forobstruction/capsule retention, incomplete smallbowel visualization, limited visualization of submu-cosal masses, and limited evaluation of the duo-denum and proximal jejunum.23

Because of the limitations of VCE, CTE may beconsidered the first-line evaluation in patientswith SBB who have suspicion of stricture (inflam-matory bowel disease, prior surgery, nonsteroidalantiinflammatory drug [NSAID] use, radiation),obstruction, or a small bowel mass (young patientsor known malignancy).24,25 A meta-analysisshowed a diagnostic yield of 40% for CTEcompared with 53% for VCE in patients withSBB.26 CTE and VCE are considered

an presents with dark red bowel movements and dizzi-trast extravasation into the transverse colon. Superse-ore delayed (C) phase demonstrates a persistent blushbolization. Patient had known angiodysplasia on the

Morrison et al796

complementary modalities; when one study isnegative, evaluation with the other should beconsidered before proceeding with DBE.25

DBE affords detailed examination of the smallbowel but is costly and time consuming. There-fore, the test is typically used in a targeted fashionafter the small bowel has been evaluated with VCEor CTE. Patients who do not receive a diagnosis orwho have recurrent bleeding after full diagnosticevaluation of the bowel (obscure GI bleeding)may benefit from repeated capsule endoscopy,Meckel scan, surgical evaluation, or observation/iron supplements.25

COMPUTED TOMOGRAPHY ENTEROGRAPHY

CTE has been used successfully for the evalua-tion of occult obscure GI bleeding for morethan a decade.27 CTE is advantageous in theworkup of occult GI bleeding because of itswide availability, relatively rapid examinationperformance, and consistent good examinationquality. CTE has benefits over endoscopic tech-niques for evaluating the entirety of the intra-abdominal GI tract, delineating postoperative GItract anatomy and evaluating extraentericstructures.Limitations of CTE include exposure to ionizing

radiation, iodinated IV contrast exposure, andneed for large-volume oral contrast administra-tion. Improvements in dose reduction techniqueshave allowed CTE radiation doses to declinesubstantially from 15 to 20 mGy to less than10 mGy per phase.28 Exposure to iodinated IVcontrast is associated with a low risk ofsevere allergic reaction (0.04%) and contrast-induced nephropathy (minimal risk with renalfunction >30 mL/min/1.73 m2).29 Large-volume

Fig. 6. Jejunal lymphoma. A 77-year-old man with historya possible small bowel mass. Axial (A) and coronal (B)mass (arrows) in the jejunum that was subsequently showthat potentially could have been missed with poor bowel

oral contrast administration may not be toleratedby patients, particularly those who are acutely ill;for some patients, placement of a nasogastriccatheter may be necessary for contrastadministration.Studies have consistently demonstrated the

ability of CTE to identify vascular, inflammatory,structural, and neoplastic conditions in the bowelresponsible for GI bleeding.27,30 Although thediagnostic yield for vascular and inflammatory le-sions may be lower than that of VCE, the tech-nique is superior for the detection of smallbowel masses and has been shown to identifyinflammation and vascular malformations notseen using VCE.24,27,30 As with all imaging studiesperformed for the evaluation of GI bleeding, thediagnostic yield of CTE is higher in patients withovert bleeding than in those with occultbleeding.23,27

CTE requires distention of the small bowel byingestion of 900 to 1350 mL of neutral oralcontrast over 30 to 60 minutes. (Please see Shan-non P. Sheedy and colleagues’ article, “CTEnterography,” in this issue for a detailed discus-sion of CTE.) A variation of enterography, enter-oclysis may be used to increase distention ofthe small bowel. Enteroclysis involves the place-ment of a naso-enteric catheter and active infu-sion of contrast agent to increase small boweldistention. Enteroclysis may have an advantagein detecting subtle strictures and pathologic con-ditions of small bowel, but it is poorly tolerated bypatients and rarely performed outside of referralcenters (Fig. 6).IV contrast administration for CTE is typically

performed using 300 to 350 mg of iodine permilliliter contrast agents and infusion ratesbetween 3 and 5 mL/s. CTE performed for the

of occult GI bleeding. Capsule endoscopy suggestedimages from a CT enteroclysis demonstrate a subtlen to be a lymphoma. Note the isodense appearancedistension with enterography.

Acute and Occult Lower Gastrointestinal Bleeding 797

evaluation of suspected small bowel bleedingrequires acquisition of multiple phases ofcontrast enhancement, although the specificphases acquired vary between institutions. Theexamination typically includes the arterial phasetimed using bolus tracking techniques and anenteric phase acquired at 50 to 60 secondsafter injection. A precontrast phase and/ora delayed phase may also be acquired depend-ing on the preferences of the radiologist.Precontrast images are helpful for identifyinghigh-attenuation ingested material, which canmimic pathology. Data demonstrating the addedbenefit of additional contrast-enhanced phasesare lacking; however, limited evidence suggestsan improvement in sensitivity and reader confi-dence when increasing an examination from 2to 3 contrast-enhanced phases.31 The use ofdual-energy data acquisition and postprocess-ing can be helpful for improving lesion detectionand limiting the radiation dose. The generationof iodine maps and low kiloelectron volt monoe-nergetic series can be helpful for makingenhancing pathology more conspicuous. Thegeneration of virtual noncontrast series maynegate the need for acquiring precontrast datasets.

Occult Gastrointestinal Bleeding: Causes andFindings

There are many potential causes of occult GIbleeding, which are listed in Table 1.32,33 In thenext section, the authors review some of themore frequently seen causes and their imagingfindings. They highlight unique or subtle findingsthat may not be well known. The discussion fo-cuses on CTE, which is the most commonly per-formed imaging test for occult bleeding. AlthoughCTE can detect lesions in the stomach and colonmissed by endoscopy, the authors focus their dis-cussion on small bowel pathology.

Table 1Common causes of occult small bowel gastrointestin

Inflammatory Neoplastic

Celiac sprueUlcerative colitisCrohn diseaseNSAID

MetastasisNeuroendocrine tumorAdenocarcinomaLymphomaGISTPancreatic heterotopiaPolyp

Abbreviations: AVM, arteriovenous malformation; GIST, gastro

Vascular lesionsVascular lesions can be classified as high-flow le-sions (arteriovenous malformation [AVM], Dieula-foy lesion) (Fig. 7), angioectasia (Fig. 8), andvenous lesions (varix, venous angioma) (Fig. 9).32

High-flow lesions are best seen on the arterialphase and may only be seen on this phase. Thepresence of an enlarged draining vein is more sug-gestive of an AVM. Angioectasias are usually bestseen on the enteric phase and appear as nodularor discoid areas of enlargement of the intramuralveins. Venous angiomas show slow progressiveenhancement and may have associatedphleboliths.

CTE has potential benefits for the evaluation ofbowel vascular malformations in addition to evalu-ation with VCE. CTE has been shown to identifyvascular lesions missed by endoscopy.27,30 Dieu-lafoy lesions may be particularly conspicuous atarterial phase enterography and can be difficultto visualize at endoscopy.

Visualization at endoscopy is limited to themucosal surface; therefore, intramural malforma-tions may only be detected at cross-sectional im-aging (Fig. 10). In addition, imaging is able tobetter localize lesions, provide a better overviewof the extent of the malformation, and reliably sur-vey the entire GI tract.

PolypsPolyps can have variable enhancement. In apoorly distended bowel, a polyp that enhancessimilar to the adjacent bowel wall may beobscured. Examinations performed with positiveoral contrast may perform better for identifyingmasses that enhance similar to the bowel wall;this should be considered when evaluating pa-tients with polyposis syndromes, such as Peutz-Jeghers syndrome.

A unique enhancement characteristic with cen-tral lower density and surrounding hyperenhance-ment is suggestive of inflammatory fibroid polyps

al bleeding

Vascular Miscellaneous

AngioectasiaAVMDieulafoy lesionVenous angioma

Meckel diverticulum

intestinal stromal tumor.

Fig. 7. High flow vascular lesion. A 71-year-old man with symptomatic anemia. Coronal MIP images from the arte-rial phase of a multiphasic CTE examination show a rapidly enhancing enlarged vessel in the ileum (arrows) sug-gestive of a high flow vascular lesion, such as a Dieulafoy lesion. The lesion was not seen on the enteric phase (notshown).

Morrison et al798

(Fig. 11). Inflammatory fibroid polyps are rare non-neoplastic lesions composed of blood vessels,connective tissue, and inflammatory cells. Theyoriginate within the submucosa, most commonlyof the stomach, with the small bowel being thesecond most common site. These lesionsfrequently become several centimeters in sizeand may present with obstruction orintussusception.

NeoplasmsThe most common primary malignancy ofthe small bowel is a neuroendocrine tumor(NET).33 NETs most commonly develop in the

Fig. 8. Angioectasia. A 68-year-old man with iron deficieenteric phase of a multiphase CTE show enlarged and nangioectasia.

ileum. With the more widespread use of CTE,small bowel NETs are being detected earlierand at a smaller size. These lesions are typicallysmall with a flat or plaquelike morphology(Fig. 12). Desmoplastic reaction associatedwith the mass may cause bending/kinking ofthe bowel loop. The small size and shape ofthe lesion may make them difficult to detect.They are best seen on the arterial or entericphases as a focus of mural hyperenhancement.When a small bowel NET is identified, it isimportant to carefully search the remainingbowel for additional lesions, as they can bemultifocal.

ncy anemia. Coronal (A) and axial (B) MIPs from theodular vessels in the jejunum (arrows) suggestive of

Fig. 9. Small bowel varix. A 68-year-old man withcirrhosis, portal hypertension, and recurrentGIbleeding.CTE demonstrates serpiginous enhancing structures inthe small bowel consistent with varices (arrow). Alsonote the varices in the anterior abdominal wall.

Fig. 11. Inflammatory fibroid polyp. A 56-year-oldman with iron deficiency anemia. CTE demonstratesan intraluminal mass (arrow) in the ileum. The masshas unique enhancement characteristics with periph-eral hyperenhancement and central low attenuationnear water density.

Acute and Occult Lower Gastrointestinal Bleeding 799

Adenocarcinoma most commonly presents asa hypoenhancing polypoid mass or an annularconstricting lesion (Fig. 13).33 Lymphoma of thesmall bowel may have a similar appearance butalso may result in aneurysmal dilation of thebowel (see Fig. 6). Lymphoma may also be

Fig. 10. Rectal vascular malformation. A 58-year-old man wdemonstrate a tangle of large vessels (white arrows) in theconsistent with a rectal vascular malformation confirmed

multifocal, associated with lymphadenopathy orlymphomatous masses in other organs. Adeno-carcinoma is more likely to obstruct thanlymphoma.

ith recent black tarry stool. Enteric phase CTE imagesrectum on coronal reformats (A) and coronal MIPs (B)on subsequent angiography (C).

Fig. 12. NET. An 82-year-old man with iron deficiencyanemia. CTE shows a small plaquelike hyperenhanc-ing lesion (arrow) in the small bowel with puckeringof the underlying wall consistent with a NET.

Morrison et al800

Gastrointestinal stromal tumors (GISTs) have var-ied appearances based on their size. When small,the masses are well circumscribed and hyperen-hancing (Fig. 14). These masses originate from amural position but typically become endoluminal orexophytic as they grow. As they grow, the enhance-ment becomes variable; necrosis, internal hemor-rhage, and ulceration may be found. Large massesmay also directly invade into adjacent structures.

Heterotopic pancreatic tissuePancreatic tissue with no anatomic continuity withthe remaining pancreas is most commonly foundin the stomach, duodenum, and proximaljejunum.34 It can also be found within Meckeldiverticula and the remainder of the ileum. Hetero-topic pancreatic tissue typically has an intramuralor serosal flat plaquelike appearance that canmimic a GIST or NET (Fig. 15).

InflammationInflammatory bowel disease is discussed ingreater detail elsewhere in this issue. Classic

Fig. 13. Jejunal adenocarcinoma. A 64-year-old man withfrom magnetic resonance enterography show a focal masssuggestive of an adenocarcinoma. Most sites prefer using Cresonance enterography can be used if there are contrain

findings of small bowel Crohn disease includeasymmetric mural hyperenhancement and bowelwall thickening in a skipping pattern. Penetrating(sinus tracts, fistula, abscess) and perianal diseasemay be present.

Nonsteroidal antiinflammatory enteropathyFocal small bowel wall injury caused by NSAIDexposure can result in thin circumferential regionsof submucosal fibrous deposition and frequentlyhave superimposed inflammation. The circumfer-ential columns of fibrosis result in short (5–10 mm long) strictures or diaphragms with super-imposed hyperenhancement, which are charac-teristic of NSAID exposure (Fig. 16). However,other causes, such as Crohn disease, may havesimilar findings. The strictures are often multipleand found in close proximity. They may bemistaken for peristaltic contractions or missedbecause of poor bowel distention, especially thosewithout superimposed inflammation. MultiphaseCT or MR imaging is helpful for identifying NSAIDstrictures by maximizing luminal distention ineach segment of the bowel.

Meckel diverticulumMeckel diverticulum occurs within 1% to 3%of thepopulation andmost (84%) are asymptomatic.35,36

Of symptomatic diverticula occurring in adults, themost common presentation is bleeding (38%).36

Fifty percent of diverticula contain ectopic tissuethat is gastric in greater than 60% and pancreaticin 16%.35

Tc-99m pertechnetate scan (Meckel scan) is auseful test for detecting ectopic gastric mucosawithin a diverticulum. The sensitivity of the test is85% to 90% in the pediatric population and 62%in patients 16 years of age and older.37 A Meckelscan may give false-positive results because ofdetection of ectopic mucosa, which may be in

melena. Coronal (A) and axial (B) T2-weighted images(arrows) in the jejunum with mild proximal dilatationTE for evaluating occult bleeding; however, magneticdications to CT.

Fig. 14. Multifocal GIST. An 82-year-old man with melena requiring multiple transfusions. CTE shows multiple hy-perenhancing lesions throughout the small bowel consistent with multiple GISTs (arrows).

Fig. 15. Ectopic pancreatic tissue. A 21-year-old woman with abdominal pain and episodes of blood in stool. CTE(A) and magnetic resonance enterography (B, C) show a mural mass (arrows) in the jejunum near the ligament ofTreitz. The lesion is somewhat flat in appearance and enhances similar to pancreatic tissue. Findings could repre-sent a GIST or NET. However, heterotopic pancreatic tissue is also in the differential diagnosis which wasconfirmed at surgery.

Acute and Occult Lower Gastrointestinal Bleeding 801

Fig. 16. NSAID diaphragms. A 64-year-old man with iron deficiency anemia. CTE demonstrates multiple focal in-flammatory strictures (arrows) consistent with NSAID enteropathy.

Morrison et al802

another location, suchasan enteric duplication cyst.Ectopic pancreatic tissue within the diverticulumcan be a source of a false-negative Meckel scan.To perform aMeckel scan, 1.85 MBq/kg Tc-99m

pertechnetate is administered.38 Pretreatment with

Fig. 17. Positive Meckel scan. A 4-year-old boy whopresents with bloody stools. On the Meckel scan,there is a focal area of increased radiotracer uptakein the right lower quadrant (arrow) consistent witha Meckel diverticulum.

pentagastrin and/or histamine receptor antago-nists may increase the rate of positive examina-tions.38 Dynamic planar images are typicallyacquired every 30 to 60 seconds for 30 minutes.Single positron emission CT can also be performedand may assist in better localization of tracer up-take. Positive examinations show a focal regionof uptake in the right lower quadrant, which ap-pears at the time the stomach is visualized(Fig. 17). Meckel diverticula can be identified byCTE. These diverticula appear as blind-endingpouches arising from the ileum. Superimposed

Fig. 18. Meckel diverticulum. A 73-year-old womanwith melena and transfusion-dependent anemia.Coronal image from a CTE shows a blind endingloop arising in the distal small bowel (asterisk) consis-tent with a Meckel diverticulum. Associated wallthickening (arrow) consistent with superimposedinflammation.

Acute and Occult Lower Gastrointestinal Bleeding 803

hyperenhancement and wall thickening suggestsuperimposed inflammation (Fig. 18).

SUMMARY

Lower GI bleeding occurs distal to the ligament ofTreitz and is an important clinical problem with avariety of causes. The bleeding can be acute, pre-senting with hematochezia, or occult, presentingwith iron deficiency anemia and/or positive fecaloccult blood test. Obscure bleeding refers to pa-tients with rebleeding after a negative endoscopicand radiologic assessment of the bowel.

The workup of lower GI bleeding includes endos-copy, CTA, CTE, nuclear scintigraphy, conventionalangiography, and surgery. The clinical presentationof patients dictates the order and urgency that thetest/intervention is performed. Radiologic assess-ment of lower GI bleeding has come to the forefrontof the workup, especially in patients with acuteongoing bleeding. Conventional angiography withembolization is now often the first-line treatment inpatients who are unstable. CTE has become afirst-line test for many patients with occult GIbleeding, especially in those whereby there is acontraindication to capsule endoscopy.

This article discusses the various imaging mo-dalities used in the workup of lower GI bleedingand includes some of the common imaging find-ings using each modality.

REFERENCES

1. Manning-Dimmitt LL, Dimmitt SG, Wilson GR. Diag-

nosis of gastrointestinal bleeding in adults. Am

Fam Physician 2005;71:1339–46.

2. Soto JA, Park SH, Fletcher JG, et al. Gastrointestinal

hemorrhage: evaluation with MDCT. Abdom Imaging

2015;40:993–1009.

3. Zimmerman HM, Curfman K. Acute gastrointestinal

bleeding. AACN Clin Issues 1997;8:449–58.

4. Gayer C, Chino A, Lucas C, et al. Acute lower

gastrointestinal bleeding in 1,112 patients admitted

to an urban emergency medical center. Surgery

2009;146(4):600–6 [discussion: 606–7].

5. Rockey DC. Gastrointestinal bleeding. In: FeldmanM,

Friedman LS, Brandt LJ, editors. Sleigenger and For-

dtran’s gastrointestinal and liver disease. 8th edition.

Philadelphia: Saunders; 2006. p. 255–99.

6. Cirocchi R, Grassi V, Cavaliere D, et al. New trends

in acute management of colonic diverticular

bleeding: a systematic review. Medicine (Baltimore)

2015;94(44):e1710.

7. Saad WE, Lippert A, Saad NE, et al. Ectopic varices:

anatomical classification, hemodynamic classifica-

tion, and hemodynamic-based management. Tech

Vasc Interv Radiol 2013;16:158–75.

8. Garcia-Blazquez V, Vicente-Bartulos A, Olavarria-

Delgado A, et al. Accuracy of CT angiography in

the diagnosis of acute gastrointestinal bleeding:

systematic review and meta-analysis. Eur Radiol

2013;23(5):1181–90.

9. Yoon W, Jeong YY, Shin SS, et al. Acute massive

gastrointestinal bleeding: detection and localization

with arterial phase multi-detector row helical CT.

Radiology 2006;239:160–7.

10. Kuhle WG, Sheiman RG. Detection of active colonic

hemorrhage with use of helical CT: findings in a

swine model. Radiology 2003;228:743–52.

11. Mayo-Smith WW, Hara AK, Mahesh M, et al. How I

do it: managing radiation dose in CT. Radiology

2014;273(3):657–72.

12. Artigas JM, Martı́ M, Soto JA, et al. Multidetector CT

angiography for acute gastrointestinal bleeding:

technique and findings. Radiographics 2013;33:

1453–70.

13. Kirchhof K, Welzel T, Mecke C, et al. Differentiation

of white, mixed, and red thrombi: value of CT in esti-

mation of the prognosis of thrombolysis—phantom

study. Radiology 2003;228:126–30.

14. Hamilton JD, Kumaravel M, Censullo ML, et al. Multi-

detector CT evaluation of active extravasation in

blunt abdominal and pelvic trauma patients. Radio-

Graphics 2008;28:1603–16.

15. Zuckier LS. Acute gastrointestinal bleeding. Semin

Nucl Med 2003;33:297–311.

16. Howarth DM. The role of nuclear medicine in the

detection of acute gastrointestinal bleeding. Semin

Nucl Med 2006;36:133–46.

17. Wang ZG, Zhang GX, Hao SH, et al. Technological

value of SPECT/CT fusion imaging for the diagnosis

of lower gastrointestinal bleeding. Genet Mol Res

2015;14(4):14947–55.

18. Emslie JT, Zarnegar K, Siegel ME, et al. Technetium-

99m-labeled red blood cell scans in the investiga-

tion of gastrointestinal bleeding. Dis Colon Rectum

1996;39(7):750–4.

19. Gunderman R, Leef J, Ong K, et al. Scintigraphic

screening prior to visceral arteriography in acute

lower gastrointestinal bleeding. J Nucl Med 1998;

39(6):1081–3.

20. Laine L. Acute and chronic gastrointestinal

bleeding. In: Feldman M, Scharschmidt BF,

Sleisenger MH, editors. Gastrointestinal and liver

disease. 6th edition. Philadelphia: Saunders; 1997.

p. 198–219.

21. Walker TG, Salazar GM, Waltman AC. Angiographic

evaluation and management of acute gastrointes-

tinal hemorrhage. World J Gastroenterol 2012;

18(11):1191–201.

22. Zuckerman GR, Prakash C, Askin MP, et al. AGA

technical review on the evaluation and management

of occult and obscure gastrointestinal bleeding.

Gastroenterology 2000;118(1):201–21.

Morrison et al804

23. Gerson LB, Fidler JL, Cave DR, et al. ACG clinical

guideline: diagnosis and management of small

bowel bleeding. Am J Gastroenterol 2015;110(9):

1265–87 [quiz: 1288].

24. Hakim FA, Alexander JA, Huprich JE, et al. CT-enter-

ography may identify small bowel tumors not de-

tected by capsule endoscopy: eight years

experience at Mayo Clinic Rochester. Dig Dis Sci

2011;56(10):2914–9.

25. Gerson LB. Small bowel bleeding: updated algo-

rithm and outcomes. Gastrointest Endosc Clin N

Am 2017;27(1):171–80.

26. Wang ZJ, Chen Q, Liu JL, et al. CT enterography in

obscure gastrointestinal bleeding: a systematic re-

view and meta-analysis. J Med Imaging Radiat On-

col 2013;57(3):263–73.

27. Huprich JE, Fletcher JG, Alexander JA, et al.

Obscure gastrointestinal bleeding: evaluation with

64-section multiphase CT enterography–initial expe-

rience. Radiology 2008;246(2):562–71.

28. Baker ME, Hara AK, Platt JF, et al. CT enterography

for Crohn’s disease: optimal technique and imaging

issues. Abdom Imaging 2015;40(5):938–52.

29. ACR. ACR manual on contrast media version 10.2.

A. C. o. D. a. C. Media. 2016.

30. Huprich JE, Fletcher JG, Fidler JL, et al. Prospec-

tive blinded comparison of wireless capsule endos-

copy and multiphase CT enterography in obscure

gastrointestinal bleeding. Radiology 2011;260(3):

744–51.

31. Thacker P, Huprich J, Barlow JB, et al. The perfor-

mance of triple-phase CT enterography compared

to single-phase and double-phase enterography

in the evaluation of obscure GI bleeding. Radiolog-

ical Society of North America 2009 Scientific

Assembly and Annual Meeting. Chicago (IL),

November 29-December 4, 2009.

32. Huprich JE, Barlow JM, Hansel SL, et al. Multi-

phase CT enterography evaluation of small-bowel

vascular lesions. AJR Am J Roentgenol 2013;

201(1):65–72.

33. McLaughlin PD, Maher MM. Primary malignant dis-

eases of the small intestine. AJR Am J Roentgenol

2013;201(1):W9–14.

34. Rezvani M, Menias C, Sandrasegaran K, et al. Het-

erotopic pancreas: histopathologic features, imag-

ing findings, and complications. Radiographics

2017;37(2):484–99.

35. Turgeon DK, Barnett JL. Meckel’s diverticulum. Am J

Gastroenterol 1990;85(7):777–81.

36. Park JJ, Wolff BG, Tollefson MK, et al. Meckel diver-

ticulum: the Mayo Clinic experience with 1476 pa-

tients (1950-2002). Ann Surg 2005;241(3):529–33.

37. Lin S, Suhocki PV, Ludwig KA, et al. Gastrointestinal

bleeding in adult patients with Meckel’s divertic-

ulum: the role of technetium 99m pertechnetate

scan. South Med J 2002;95(11):1338–41.

38. Grady E. Gastrointestinal bleeding scintigraphy

in the early 21st century. J Nucl Med 2016;57(2):

252–9.